Summary of Argument Structure Seminar on April 29, 2005

Idioms and passives

The two domains story raises the question of exactly where the boundary is. It is clearly below tense and modality and clearly above the process and result meanings associated with verbs; on Friday we attempted to localize it more narrowly by examining the lowest of the various verbal inflections in English, the passive. The meaning and structure of the eventive passive is so consistent and transparent in English that it seems to be a good candidate for something which is outside the domain boundary. The adjectival passive, on the other hand, admits of a few idiosyncracies which suggest that it might be below the boundary. So the prediction is that there should be adjectival passive idioms but no eventive passive idioms.

English provides a number of idioms which appear only in the passive. They can be divided into two groups. First, there are those for which the promotable internal argument is specified as part of the idiom:

[cast the die] -- > the die is cast, the die will be cast, the die seems to be cast; 'the decision is made'

[run the race] --> the race is run, the race will be run, the race seems to be run; 'the opportunity is over'

[made it] --> We had it made, John has got it made, They seem to have it made; 'have a good, secure position'

[said and done all when] --> When all is said and done; When all was said and done; When all has been said and done; 'despite obstacles'

I exclude the following since it has no "free" parts, i.e. it is an adage:

[not built Rome in a day] --> Rome wasn't built in a day; ??Rome must not have been built in a day; ??Rome is sure not to have been built in a day

The above examples are arguably all adjectival. Second, there are some idioms for which the gap is a free position:

[made x for each other] 'ideally suited as a pair'

[cast x in stone] 'permanent, inflexible' (of a decision or rule)

[caught x short] 'out of' (e.g. money)

[taken x aback] 'astonished'

[written x on water] 'has no impact'

Again, these usually pattern as adjectival (He seems caught short, That rule doesn't sound cast in stone, They look made for each other), with exceptions discussed below.

A particularly strange one is the following:

[fit to be tied x] 'furious'

This seems to involve a passive under an adjective, which might be a problem for my generalization that idioms should not span a domain boundary. Interestingly, some speakers don't actually have this syntax for the adjective "fit"; for them, it is a negative polarity item (NPI) which takes an infinitival with an arbitrary PRO subject; thus, they can say

This isn't fit to eat; These aren't fit to wear; That doesn't seem fit to serve to guests

But not

%?This is fit to eat; %?These are fit to wear; %?That seems fit to serve to guests

(because of NPI) and not

%*This isn't fit to be eaten; %*These aren't fit to be worn; %*That doesn't seem fit to be served to guests

(because of arbitrary PRO). I include percent signs before the judgments since other speakers allow all of these sentences. However, those speakers don't have the idiom "fit to be tied"; so an interesting possibility is the following: if you have a syntax that allows a passive under "fit," then you will parse "fit to be tied" as something too big to be an idiom, and interpret it literally (like the second group of speakers does); but if your syntax doesn't allow a passive under "fit" (like the first group), then you are forced to parse "fit to be tied" as something else, and allow it to be small and an idiom. This is pretty speculative and should be investigated with additional speakers (so far each group consists of one).

The most damaging example encountered so far is:

[hoist(ed) x with/by/on x's own petard]: He was hoisted with his own petard; I was hoisted by my own petard; You will be hoisted on your own petard 'damaged by own scurrilousness'

(Gillian has "hoist," as in Shakespeare's Hamlet, whence the original "For 'tis the sport to see the enginer hoist with his owne petar")

Incidentally, for some speakers, "petard" becomes plural if the subject is plural:

They were hoisted with their own petards

Even though there is no word "petard" in the language other than in this idiom. This suggests some sort of compositionality which is not readily available to introspection.

The idiom is damaging to the hypothesis because it is not convincingly adjectival; compare

The decision was cast in stone when the candidate arrived. ENTAILS It had been decided before the candidate arrived (adjectival)

He was hoisted by his own petard when he drank the poisoned wine. CONSISTENT WITH He was undone by his own deed at the very moment he drank the wine. (eventive)

So, in conclusion it seems that "hoist with x's own petard" is a counterexample to the story so far.

Peter